Meta eyes $10 billion investment in Scale AI amid growing global competition

New Delhi: Meta is reportedly in discussions to invest more than $10 billion in Scale AI, a data-labelling company that has become an increasingly important player in the artificial intelligence (AI) ecosystem. If finalised, it could be one of the largest private funding deals to date. In 2024, Scale AI was valued at approximately $14 billion.

The potential investment comes amid heightened global competition in AI. Just three months after Chinese developer DeepSeek launched a model rivalling top American systems, Scale AI’s 28-year-old CEO Alexandr Wang testified before US lawmakers, urging them to act swiftly to retain technological leadership.

Appearing at a Capitol Hill hearing in April, Wang advocated for a national AI data reserve, increased power capacity for data centres, and a more unified regulatory approach across states. “It’s good to see you again here in Washington. You’re becoming a regular up here,” said Representative Neal Dunn of Florida.

Though less well-known publicly than OpenAI’s Sam Altman, Wang has gained influence in both technology and policy circles. His company provides labelled data used to train AI models for firms such as Meta and OpenAI. It also supports businesses in building bespoke AI tools and applications. Recently, Scale AI has begun working with highly qualified professionals, including PhDs and healthcare specialists, to enhance model complexity and performance.

In AI development, data stands alongside computing chips and expert talent as a foundational element. Scale AI has built a significant position in the data segment of the industry.

Founded nearly a decade ago, Scale AI’s growth trajectory resembles that of OpenAI. The two CEOs are reported to have briefly shared a residence and have both spoken before Congress. Like OpenAI, Scale AI has attracted major tech industry funding.

Originally, the company focused on labelling visual data for autonomous driving technologies. It has since shifted towards preparing vast textual datasets for training large language models (LLMs) like ChatGPT. These models depend heavily on clean, well-labelled data to learn tasks such as conversation, content generation, and reasoning.

However, the company’s heavy use of offshore contractors has faced criticism. In countries such as Kenya and the Philippines, workers have reportedly dealt with distressing content while earning modest pay. Wang has previously defended the firm’s compensation, stating in 2019 that its contractors were paid in the “60th to 70th percentile of wages in their geography”.

According to spokesperson Joe Osborne, the US Department of Labor recently closed an investigation into the company’s labour practices.

As synthetic data begins to gain traction in AI training, Scale AI’s conventional labelling services may face challenges. Yet the industry continues to grapple with a shortage of high-quality, real-world data, particularly for building advanced models capable of human-like reasoning.

To meet this demand, Scale AI has pivoted to employing more specialised contributors. These experts participate in reinforcement learning processes—evaluating AI-generated outputs and helping guide improvements. As of early 2025, around 12% of the firm’s contributors held doctorates, while over 40% had a master’s degree, law degree, or MBA.

This work increasingly targets complex and regulated fields, such as medical diagnostics and legal compliance. In one example, Scale AI has focused on improving models’ ability to answer tax law queries, which can vary significantly across countries and even individual states.

The firm reported around $870 million in revenue for 2024 and expects to cross $2 billion in 2025. Since DeepSeek’s debut, there has been rising demand for more refined and capable AI models, further boosting Scale AI’s relevance.

In parallel, the company has strengthened ties with the US government, particularly through defence contracts. Wang, who has voiced concerns about China’s role in AI, has developed relationships with lawmakers focused on national security. Michael Kratsios, a former Scale AI executive, now serves as a senior technology advisor to President Donald Trump.

For Meta, a deeper alliance with Scale AI could bolster its position in the AI arms race and offer opportunities in defence technology. For Wang, the move also brings a personal resonance. Reflecting on the company’s founding, he once said, “I rattled off some silly answer about being inspired by The Social Network,” the film that chronicles the early days of Facebook.